At the 2020 waic (World Artificial Intelligence Conference), Li Kaifu, chairman of Innovation workshop, put forward the viewpoint of “reconstructing China’s economy with technology from AI + to + AI”.

Li Kaifu mentioned that in the past few years, AI has undergone a particularly large transformation from technology driven to commercial driven. The industrialization and commercialization speed is getting faster and faster. There are four important reasons:

1. Progress of software tools;

2. Hardware acceleration makes it easier to use;

3. There are many cloud and other technologies;

4. More and more talents have been trained, and there are a large number of AI talents

With these four important factors, AI entrepreneurship has been pushed from AI + to + AI era.

We now see that from 2012 to today, the Internet must prevail. Finally, the two are integrated, so AI + and + AI will eventually integrate, but will this process be the same?

Looking back at the AI + era, early AI companies focused on technology and started businesses with talented scientists as the core. There are very few such companies. After all, there are very few AI talents, so these AI talents are supported by capital and pursue to become the first AI company.

About four or five years ago, we saw that AI + did a lot of good work, but more and more talents who understand AI and tools are becoming more and more popular, so more traditional companies began to think about how they should integrate into AI, so we officially entered the + AI era.

What is the difference between AI + and + AI?

AI +: it takes AI as the core and AI engineers and scientists as the core to look for business opportunities;

+AI: it is an AI application that generates value for traditional applications, and then must cooperate with traditional companies, or even dominated by traditional companies, and traditional companies will look for AI companies to help them.

Of course, in about five years, we believe that AI will enter the next stage. AI is ubiquitous, and AI applications will become simpler and simpler. Traditional companies can hire engineers, and they can introduce AI into their companies with simpler and more grounded methods, just like today’s it state. So to put it more appropriately,

We can see these examples. AI + company mainly focused on voice, vision and chip in the early stage. For example, Wen yuanzhixing cooperated with Guangzhou taxi company. It can be seen that AI companies exist, but AI companies are to assist a traditional company to achieve business indicators, which is the difference between + AI and AI +.

What is the difference between AI + and + AI?

PwC’s prediction tells us that AI will bring a value of RMB 1 trillion to the world in 2030. This value is mainly reflected in traditional enterprises.

Why does AI become so important in the next stage of traditional enterprises?

First, the volume of traditional industries is very large. A bank or a car manufacturing company has a large volume. If AI can help improve its efficiency by 3% and 5%, it will produce great value;

Second: the threshold of traditional companies is very high, but the threshold of AI is getting lower and lower. It is possible to think that the high-tech threshold is the highest. As mentioned earlier, AI is becoming easier and easier, so the AI threshold is decreasing. The threshold of traditional enterprises is very high. In short, it is very difficult for AI companies to be a bank, but it is relatively easy for a bank to integrate into AI;

Third: traditional industries have the important role of industrial chain, so the connection between upstream and downstream will enable traditional industries to form such a necessary function.

Finally, we found that although artificial intelligence is very powerful, its popularity is still limited. No artificial intelligence has become a platform that can be used directly. Every traditional enterprise has considerable customization. For example, unique data needs to be collected and cleaned up. For example, manufacturing companies even need to add more sensors. These are all customization needs, not a set of products that can be used by traditional companies.

The Internet is very different from AI. The Internet contacts every user to form a huge platform. AI is more of a great technology, which will empower the existing platform.

So whether there will be a platform for AI, Li Kaifu said: not so optimistic

The value generated by AI itself is as large as that of the Internet. According to these reasons, AI + will continue to be valuable, but + AI has a larger volume and makes a greater contribution to the social economy.

What kind of traditional enterprises need to consider + AI?

In fact, AI is valuable in every enterprise, but traditional industries often have many curing habits. Not every company is suitable to join AI today. Here are three sets of suggestions:

First, the company is still growing. Now it needs to expand or reduce costs. There is a business demand such as + AI.

Second, the company’s own data is good enough, with enough data, enough structured data, and data related to business indicators. If it can integrate AI data, it can create a real business value.

Third, traditional companies are sometimes solidified, old and difficult to change. They are more suitable for AI. The company will be a visionary CEO. The company has a good enough culture and is willing to transform itself. If such companies do AI today, there are four ways to make AI produce value. First, save money with AI, simply replace one link with AI, transform several important processes of the company with AI, and reconstruct the whole industry with AI. The first two or three will see more examples. This is a classic traditional enterprise + AI.

+Four cases of AI

1. Single link cost saving

The first one is to use AI vision for quality inspection to see if the produced clothes have color or size problems;

Second, we can use AI to see whether the display of goods gives us enough space for each product, and how much stock is left to help retail stores do AI intelligent empowerment;

Third, use AI to detect the production process. Here is a motor to see if there are defects in it. Finally, use intelligent method to see this process in the molten iron process.

2. Single link optimization enabling

Can the loan finance industry use AI to replace a single link? When lending, AI is used to decide whether to lend money to a person, and a massive data experiment is used. Such an experiment is also conducted in the innovation factory engineering institute and a very large loan company. AI is used to replace people as the judgment of loan, which reduces the default rate by about 14% and saves huge expenses.

3. Complex intelligent operation

The more complex process enabling is the aspect of intelligent operation. If AI is used to help the store predict each commodity sold and how many copies are sold in each store, it can make a better prediction, so that it can reduce the possibility of stock shortage, that is, increase its business and make its inventory more intelligent. After this process is completed, we will work in the whole logistics, warehousing In terms of manufacturing, we can connect with AI one by one. Slowly, we will be very smart to know how much each product should be produced, how much to store, where to store it, when to sell it out, when to replenish it, and even where to put it in the mall to optimize the final sales.

4. Reconstruction industry

It is possible to reconstruct the rules of the whole industry. In fact, medical treatment is a great opportunity. Let’s talk more deeply about using AI pharmaceutical. Investing in a company is to find the most suitable small molecules chemically, and add the order of AI for various pharmaceuticals to see which drugs and diseases can produce the greatest economic value and help more people recover their health, In addition, we can have a higher probability and pass the clinical experiment. In this process, we found that if everything goes well, it is possible to invent new drugs, which can be 4-5 times faster than in the past. In fact, this has successfully reconstructed the medical industry.

Innovation factory has invested in more than 40 cases in artificial intelligence. Recently, it has found that more and more AI companies must be able to land and connect with traditional companies in the future. Moreover, its value lies in its empowerment + AI, that is, a good AI case creates the greatest value. It is almost impossible to have a traditional company, He uses the traditional foundation to use AI for empowerment and value creation.

+Problems and trends faced by AI

Today, China is facing a very important moment, that is, our traditional industries need to reduce costs and increase efficiency, and face several major problems.

First, labor costs are becoming more and more expensive

As a large manufacturing country in the world, the labor cost is becoming more and more expensive, and the productivity and efficiency are still quite insufficient. If we actually look at the retail and production links, we find that there are various aspects that need to improve the efficiency, such as the productivity in the factory, or although we bring so much consumer efficiency than Taobao and pinduoduo, However, many things in the background still lack efficiency. In addition, there are many offline businesses that need to improve efficiency. There are seven million mom and pop stores. Finally, there are many industries, including education and medical industry. Its efficiency needs to be improved. Artificial intelligence can be used to improve efficiency, so from the development of the whole industry, We can see that the great value in the past 10-20 years comes from front-end innovation, which enables us to surf the Internet in many industries, make it more convenient to use app, create a lot of value and generate a lot of data. In the next 10 years, we will see that more traditional industries, manufacturing, medical treatment and education, are facing their next stage of improvement. It is to make targeted changes to improve their efficiency in the current stage, use AI to improve empowerment, and use the four methods just mentioned to create new value. This will be the greatest opportunity for the improvement of the whole economy and the best investment direction for us to see when we invest.

Second, the trend of AI is promoted under the epidemic situation

There are two major factors that make us very optimistic about + AI. The first factor is that although today’s epidemic is a great disaster to the world, it has actually changed our usage habits, changed more offline businesses to online, turned non data into data, and generated more data can also make better AI. We see that in medical treatment, office More and more industries such as education are taking the data route. I believe + AI will accelerate the landing of AI in these industries because of the epidemic.

Third: new infrastructure

To do a good job of industrial AI and let traditional enterprises embrace AI, we need to have a very good foundation in all aspects of computing, communication and data before we can run. In the past, if there was no good infrastructure, roads and urban construction, we could not give full play to its value in building cars. Today, if we want to make good use of AI in various traditional industries, such as data center, 5g, IOT Big data and AI are very important, and these links will be combined.

“Therefore, I deeply believe that + AI can achieve data, it and cloud in one step under the new infrastructure, which will play an important role in improving China’s economy,” Li Kaifu said.

        Editor in charge: Tzh

Leave a Reply

Your email address will not be published. Required fields are marked *